Graphene Oxides in Water: Correlating ... - ACS Publications

Jun 1, 2016 - through a small opening under applied pressure of 14 psi using nitrogen as the carrier gas. The liquid/gas jet ... software PHI Multipak...
2 downloads 20 Views 2MB Size
Subscriber access provided by UNIV OF NEBRASKA - LINCOLN

Article

Graphene Oxides in Water: Correlating Morphology and Surface Chemistry with Aggregation Behavior Yi Jiang, Ramesh Raliya, John D. Fortner, and Pratim Biswas Environ. Sci. Technol., Just Accepted Manuscript • Publication Date (Web): 01 Jun 2016 Downloaded from http://pubs.acs.org on June 1, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 27

Environmental Science & Technology

Graphene Oxides in Water: Correlating Morphology and Surface Chemistry with Aggregation Behavior

Yi Jiang, Ramesh Raliya, John D. Fortner*, and Pratim Biswas*

Department of Energy, Environmental, and Chemical Engineering, Washington University in St. Louis. St. Louis, Missouri 63130, United States

Submitted to Environmental Science & Technology February, 2016

*To whom correspondence should be addressed: Pratim Biswas.: Tel: +1-314-935-5548; Fax: +1-314-935-5464; E-mail: [email protected] John D. Fortner: Tel: +1-314-935-9293; Fax: +1-314-935-5464; Email: [email protected]

ACS Paragon Plus Environment

Environmental Science & Technology

Page 2 of 27

1

Abstract

2

Aqueous aggregation processes can significantly impact function, effective toxicity, environmental

3

transport, and ultimate fate of advanced nanoscale materials, including graphene and graphene

4

oxide (GO). In this work, we have synthesized flat graphene oxide (GO) and five physically

5

crumpled GOs (CGO, with different degrees of thermal reduction, and thus oxygen functionality)

6

using an aerosol method, and characterized the evolution of surface chemistry and morphology

7

using a suite of spectroscopic (UV-vis, FTIR, XPS) and microscopic (AFM, SEM, and TEM)

8

techniques. For each of these materials, critical coagulation concentrations (CCC) were determined

9

for NaCl, CaCl2, and MgCl2 electrolytes. The CCCs were correlated with material ζ-potentials (R2 =

10

0.94-0.99), which were observed to be mathematically consistent with classic DLVO theory. We

11

further correlated CCC values with CGO chemical properties including C/O ratios, carboxyl group

12

concentrations, and C-C fractions. For all cases, edge-based carboxyl functional groups are highly

13

correlated to observed CCC values (R2 = 0.89-0.95). Observations support the deprotonation of

14

carboxyl groups with low acid dissociation constants (pKa) as the main contributors to ζ-potentials

15

and thus material aqueous stability. We also observe CCC values to significantly increase (by 18-

16

80%) when GO is physically crumpled as CGO. Taken together, the findings from both physical

17

and chemical analyses clearly indicate that both GO shape and surface functionality are critical to

18

consider with regard to understanding fundamental material behavior in water.

19 20 21 22 23 24 1 ACS Paragon Plus Environment

Page 3 of 27

25

Environmental Science & Technology

Introduction Graphene oxide (GO) shares the one-atom-thick planar sheet with sp2-bonded carbon structural

26 27

framework as graphene, but with oxygen-containing functional groups which include basal

28

hydroxyl and epoxy, and edge-associated carbonyl and carboxyl groups.1, 2 Based on unique

29

material properties, GO has been widely studied for a number of advanced applications including

30

energy conversion and storage,3 enhanced catalysis,4, 5 antimicrobial,6, 7 sorption,8 and separations,9,

31

10

32

widely expected to grow significantly in the coming decade.11 Upon commercial production and

33

application, and thus environmental exposure, concerns have been raised regarding the potential

34

biological effects, including to human health, as GO has been observed to be cytotoxic to

35

mammalian cells and bacteria.12-15 Interestingly, both material functionality and exposure, which is

36

critical with respect to biological response, are a function of CGO aggregation state/behavior in

37

water.16, 17

among other technologies. Further, the production and application of GO-based materials are

38

Despite being an essential component for quantitative material behavior models considering

39

both application and potential negative implications, fundamental description of GO aggregation

40

behaviors in water is currently incomplete. GO materials typically vary in nature, due to the random

41

functionalization for each layer and variations in physical structure (such as molecular weight,

42

shape, defects).18, 19 Residual oxygen moieties can, depending on methods and degree of reduction,

43

differ significantly from a few to dozens of percent in terms of atomic ratio.19 Further, 2D GO can

44

be physically modified, resulting in 3D structures, such as crumpled paper ball-like spheres20, 21 and

45

corrugated (wrinkled) surfaces.22

46

Initial reports on the aqueous aggregation and transport behavior of graphene materials have

47

focused on pristine, flat GO.23, 24 Compared to pristine GO, aqueous stability of GO derivatives (e.g.,

48

GO with different degrees of reduction and morphological transformation) can differ significantly 2 ACS Paragon Plus Environment

Environmental Science & Technology

49

under similar aqueous chemistries. Due to the complexity of possible chemical and physical

50

variations, a quantitative understanding on how such intrinsic structures and properties affect GO

51

aqueous stabilities is needed. However, such understanding remains challenging due to the lack of

52

convenient, yet consistent (reproducible) control and characterization of physical and chemical

53

properties.

54

Page 4 of 27

We have previously demonstrated GO morphological control by physically crumpling 2D flat

55

GO into 3D crumpled structures (termed as CGO) via a furnace aerosol reactor (FuAR) method,

56

using GO coupons as the starting material.21 The method utilizes capillary compression induced by

57

rapid evaporation of the aerosol droplets to effectively crumple flat GO. Furthermore, the surface

58

chemistries (degree of thermal reduction) can be tuned by precisely varying the furnace temperature

59

(200-800 °C) while maintaining the crumpled structure. In this work, we synthesized GO using the

60

modified Hummer’s method,25 and then five subsequent CGO materials, each with different degrees

61

of reduction (crumpling GO at different furnace temperatures from 200 to 800 °C, all with the same

62

starting coupon structures). Based on extensive characterization and aggregation kinetic results, we

63

have correlated critical coagulation concentration (CCC) values for three ionic systems (NaCl,

64

CaCl2, and MgCl2) with physical and chemical properties of GO/CGO (ζ-potentials, C/O ratios,

65

carboxyl, and C-C fractions). We also observe an increase of CCC values for CGO materials when

66

compared to flat analogues (comparing GO and CGO with same surface chemistry). This is the first

67

report that provides a quantitative description of GO aggregation as a function of both morphology

68

and surface chemistry.

69 70

Materials and Methods

71

Synthesis of GO/CGOs. GO was synthesized using the modified Hummer’s method25 and was

72

detailed in our previous work.21 Functional groups such as epoxy, hydroxyl, and carboxyl adorn the 3 ACS Paragon Plus Environment

Page 5 of 27

Environmental Science & Technology

73

surface of GO to render it dispersible in polar solvents including water.26 Crumpled GO particles

74

(CGO) were synthesized by a furnace aerosol reactor (FuAR) method using GO as the starting

75

material.21 35 mL of ~50 mg/L GO solution was placed in a six-jet Collison nebulizer jar (BGI

76

Incorporated), and the pressure nebulizer produced water droplets by forcing the solution through a

77

small opening under applied pressure of 14 psi using nitrogen as the carrier gas. The liquid/gas jet

78

was impacted against the inside wall of the jar to remove larger fraction of the droplets, and the size

79

of the outflow water droplets was mainly micrometer-sized (2-4 µm) as previously measured by an

80

aerosol particle sizer (APS).21 The water droplets containing GO sheets were then delivered by

81

nitrogen gas into an alumina reactor (1 m × 25 mm ID) maintained at predetermined temperatures

82

(from room temperature to 1000 °C) to heat it for several seconds. The flow rate is generally

83

operated at 12.4 L/min (nebulizer pressure 14 psi (96.53 kPa)), resulting in ~1.6 s residence time.

84

The formed CGO nanoparticles were finally collected at the end stream of the reactor, weighed and

85

dispersed in water to get 200 mg/L dispersion.

86

Characterization of GO/CGOs. The morphology and size of the GO/CGO samples were

87

examined by transmission electron microscopy (TEM, TecnaiTM Spirit, FEI Co.) and field

88

emission scanning electron microscopy (FESEM, NOVA NanoSEM 230, FEI Co.). For GO SEM

89

imaging, samples were sputter coated with gold for 90 s (Headway PWM32-PS-CB15PL). The size

90

and thickness of GO were also measured using atomic force microscopy (AFM, Veeco Nanoman).

91

The optical properties of GO/CGO aqueous dispersions (20 mg/L) were measured by using a UV-

92

vis spectrophotometer (Varian Bio 50). Surface chemistry information regarding molecular bond

93

and functionality were obtained with fourier transform infrared spectrometer (FTIR, Nicolette

94

Nexus 470) and X-ray photoelectron spectroscopy (XPS, PHI 5000 VersaProbe II equipped with

95

monochromatic Al Kα (1486.6 eV) X-ray source). The XPS peaks were fitted to a mixed function

96

having 80% Gaussian and 20% Lorentzian characters using the software PHI Multipak, after 4 ACS Paragon Plus Environment

Environmental Science & Technology

Page 6 of 27

97

performing a Shirley background subtraction. In the fitting procedure, the FWHM values were fixed

98

at 1.2 ± 0.2 eV for all peaks, and the peak positions were constrained within 0.2 eV deviated from

99

the assigned position. Calibration was carried out by alignment of the spectra with reference to the

100

C 1s line at 284.8 eV associated with graphitic carbon. At least three measurements were performed

101

at different samples (or locations). ζ-potential and hydrodynamic diameter (Dh) (in 40 mg/L

102

aqueous solution) were measured with a ZetaSizer Nano ZS (Malvern Instruments, Worcestershire).

103

Aggregation Kinetics of GO/CGOs. The early-stage aggregation kinetics of GO/CGOs were

104

assessed by the initial rate of change of the Dh with time t. In the early aggregation stage, the initial

105

rate constant (ka) is proportional to the initial rate of increase in Dh and inversely proportional to the

106

initial (primary) nanoparticle concentration in the suspension (N0) (Eqn. 1).27

107

 =   ( )

108







(1)

→

The attachment efficiency (ɑ) (also known as the inverse stability ratio) at different electrolyte

109

concentrations was calculated by normalizing the aggregation rate constant obtained to the rate

110

constant obtained under favorable (non-repulsive, fast) conditions (ka,fast) (Eqn. 2).27

111

112

=

 

=

    ( )    →     ( ) ,  →,



(2)

The initial rate of increase in Dh was measured by time-resolved dynamic light scattering (TR-

113

DLS) (Malvern ZetaSizer Nano ZS). Equal volumes (500 µL) of GO/CGO dispersion and

114

electrolyte solution (NaCl, CaCl2, and MgCl2 with different ionic strength) were mixed to reach the

115

desired concentrations (GO/CGO: 40 mg/L; NaCl: 0 - 250 mM; CaCl2 and MgCl2: 0 - 50 mM).

116

Previous studies on aggregation of carbon nanomaterials were usually conducted with a pH between

117

5 and 6 (e.g., graphene oxide, pH 5.523 and C60, pH 5.227) , and thus the pH of the GO/CGO

118

dispersion was adjusted to 6.0 ± 0.3 (with 0.02 mM-0.5 mM NaOH and/or HCl) for comparison.

119

The DLS glass cuvette was quickly vortexed and placed in the instrument for measurement. The 5 ACS Paragon Plus Environment

Page 7 of 27

Environmental Science & Technology

120

rate was calculated for the initial stage defined as the period between t = 0 to the time when Dh

121

reaches 1.30Dh0.27, 28 Since GO/CGO concentration remained identical, α was then determined to be

122

the ratio of the initial rate of change of Dh in the reaction-limited regime over that in the diffusion-

123

limited regime. Critical coagulation concentrations (CCC) were determined from the intersection of

124

extrapolated lines through the diffusion and reaction limited regimes.

125 126

Results and Discussion

127

Materials Characterization. The aerosolized droplets then undergo solvent evaporation, and

128

capillary compression induced by rapid evaporation can effectively crumple flat GO. The

129

magnitude of the compression, which has been correlated to the evaporation rate of solvent, was

130

identified as the critical factor determining the morphology as well as the size of the dry CGO

131

particles.21 While being crumpled, simultaneous thermal reduction leads to partial removal of

132

surface functional groups, and restoration of aromatic carbon regions.3, 5 Different degrees of

133

reduction, while producing the same crumpled morphology, were achieved by varying the furnace

134

temperatures (e.g., 200 to 800 °C), as revealed in our previous work.21

135

As-synthesized GO/CGO samples were characterized by TEM (Figure 1a-c), FESEM (Figure

136

S1), AFM (Figure 1d and Figure S1g), UV-vis (Figure S3), XPS (Figure 1g and Figure S4), and

137

FTIR (Figure S5) for detailed size, morphology and surface chemistry information. Figure 1a-c

138

compares the morphologies of GO and representative CGO materials prepared at furnace

139

temperatures of 200 and 400 °C (CGO reduced at 200 °C, denoted as CGO-200 hereafter; same

140

denotation applied to other CGOs). Flat GO coupons are observed with sizes ranging from a few

141

hundred nm to more than 1 µm (Figure 1a and S1a), and AFM examination of the height reveals

142

that most GO sheets are single layer or double layers (h < 2 nm, Figure 1d), consistent with

143

previous reports of GO synthesized by the modified Hummer’s method.23, 29 As-synthesized CGO 6 ACS Paragon Plus Environment

Environmental Science & Technology

144

particles, with crumpled morphology and sharp ridges (Figure 1b and c, and S1b-f), have a fractal

145

dimension of ~2.5, similar to that of crumpled paper balls.30 The fractal dimension (f) relates the

146

particle mass (m) with the diameter of crumpled ball-like structures (d) through a power law

147

expression (m ~ df). While crumpled particles may have same fractal dimension, they could have

148

different diameters depending on the degree of applied confinement force.21, 31 We have analyzed

149

the size distribution of GO and CGOs from AFM (GO) and TEM (CGO) images using software

150

ImageJ. For each measurement, approximately 150 particles were counted. For GO and CGO-200,

151

they have a relatively wide size distribution from 100 to 500 nm, but for other CGOs, they have a

152

similar and narrower size distribution, with about 80% between 100 and 300 nm (Figure 1e). This

153

trend of size change is consistent with our previous study,21 showing higher evaporation rate under

154

higher furnace temperature leading to larger confinement force and thus smaller particle size. The

155

hydrodynamic diameters of GO and CGOs are in the range of 200-350 nm as measured by dynamic

156

light scattering (DLS) (Figure 1f). It should be noted that in DLS measurement, a non-spherical

157

particle is treated as a sphere that has the same average translational diffusion coefficient as the

158

particle being measured. For flat GO sheets, the DLS measurement has been shown to

159

underestimate the real particle size,32 which is consistent with our observations. For GO and CGOs

160

synthesized below 600 °C, ζ-potentials are below -40 mV; while for CGO-600 and CGO-800, due

161

to significant thermal reduction, ζ-potentials increase to -20 mV; values higher than -30 mV are

162

usually considered as threshold for colloidal stability in water (Figure 1d).26

163

The color of the suspended samples gradually changes from brown (GO and CGO-200) to

164

black (CGOs synthesized at ≥ 400 °C) due to thermal reduction (Figure 1a-c insets and Figure S2),

165

suggesting progressive restoration of the π network within the carbon structure.33, 34 Two

166

characteristic absorption peaks of GO are observed (Figure S3) at 230 nm and 300 nm for the π-π*

167

C=C transition band and the n-π* C=O transition band, respectively.35 Upon reduction, the major 7 ACS Paragon Plus Environment

Page 8 of 27

Page 9 of 27

Environmental Science & Technology

168

absorption peak (230 nm) is observed to be red shifted (to 270 nm, typical absorption peak of

169

graphene), and the absorption in the whole spectral region (> 230 nm) increases with the degree of

170

redution, indicating partial restoration of electronic conjugation.26

171

XPS was employed to evaluate the evolution of oxygenated functionality during the thermal

172

reduction process. Survey spectra show C/O ratio of GO to be 1.9 ± 0.1, which is typical of GO

173

synthesized by the modified Hummer’s method (~ 2.0).36 While it does not change for CGO-200

174

(2.0), the C/O ratio increases to 3.2 ± 0.1 for CGO-400, and to 5.0 ± 0.7 for CGO-800. Further,

175

high-resolution C 1s spectrum of GO exhibit well-defined, multi-peak formations, indicating

176

extensive material oxidation (Figure S4a). When crumpled, CGO-200 has a very similar C 1s

177

spectrum, due to preservation of surface chemistry as GO (Figure S4b) (also evidenced by the

178

brown color of the solution (Figure S2), C/O ratios from XPS survey spectra (1.9 vs. 2.0), and

179

similar FTIR spectra, which is shown in Figure S5). Upon further thermal reduction, the peak

180

symbolizing lower oxidation state (C-C) becomes prominent, while peaks of higher oxidation states

181

decreases (CGO-400, 500, 600 and 800, Figure S4c-f).

182

C 1s spectra were deconvoluted and analyzed for carbon oxidation states (Figure 1g and Figure

183

S4a-f). All peak positions and FWHM were strictly constrained with ± 0.2 eV deviation. The

184

FWHM values were fixed at 1.2 ± 0.2 eV for all major peaks, and the peak positions were

185

constrained within 0.2 eV from the assigned position. The detailed peak position and FWHM

186

information was provided in the supporting information (Figure S4g and h). The five most

187

commonly accounted components, including the C-C (284.8 eV), C-OH (286.2 eV, 1-1.5 eV shift to

188

higher binding energy (BE)), C-O-C (287.1 eV, higher BE compared to C-OH group), C=O (287.7

189

eV, 2.5-3 eV shift to higher BE) and COOH (288.8 eV, 4-4.5 eV shift to higher BE)

190

functionalities,36, 37 were identified.

8 ACS Paragon Plus Environment

Environmental Science & Technology

191

Page 10 of 27

The relative ratio of each component to the C 1s peak is illustrated in Figure 1e. The C-C area

192

ratios increase gradually from ~41 ± 4 % of GO to 75 ± 4 % of CGO-800, with the range being

193

similar to a previous study.36 This trend coincides with the change in C/O ratio and restoration of

194

aromatic regions. Consistent with previous reports,36 the total contribution of C-O (including C-OH

195

and C-O-C) groups remains almost constant for CGO-200 (compared to GO), indicating the

196

temperature and short residence time in the furnace was insufficient to significantly affect

197

occurrence of these functional groups. However, above 200 °C, C-O-C groups are observed to

198

decrease dramatically, while the relative C-OH peak area ratio increases (Figure 1f). The C-OH

199

groups first increase from 200 °C, then decrease above 500 °C, thereby remaining stable to 800 °C,

200

which is also similar to reports by others.36 This increase is likely due to the transformation of C-O-

201

C to C-OH groups. For GO synthesized by the Hummer’s method, which typically has a C/O ratio

202

of ~2, carboxyl groups were identified to have a contribution of around 6% (to relative carbon

203

oxidation state).38-40 In our analysis, the carboxyl fraction gradually (288.8 eV) decreases as a

204

function of furnace temperature from 4.9 ± 0.6% of GO to 4.6 ± 0.3% of CGO-400 and 3.0 ± 0.6%

205

of CGO-800 (Figure 1e).

206

In general, FTIR measurements agree with XPS analysis. For as-synthesized GO, a mixture of

207

oxygen-based functional moieties including C-O (phenolic/epoxy/carboxyl), C=C (aromatic), C=O

208

(carbonyl), and -OH (hydroxyl) stretches are observed (Figure S5).5, 39, 41 Broad and strong OH

209

bands at ~3200 cm-1 and 1620 cm-1 for GO and CGO-200 are indicative of bound water

210

molecules,39 revealing high hygroscopicity (hydrophilicity, and maintaining of surface functional

211

groups). These bands decrease dramatically for samples synthesized at 400 °C and above, likely by

212

restoration of the basal aromatic fractions. Further, the 1580 cm-1 adsorption, which corresponds to

213

aromatic C=C band, is observed to be prominent for CGO-400, CGO-500, CGO-600, and CGO-800

214

materials. The evolution of bands at ~1730 cm-1 (carbonyl) and ~1425 cm-1 (C-O, carboxyl) also 9 ACS Paragon Plus Environment

Page 11 of 27

Environmental Science & Technology

215

indicates carboxyl group reduction (Figure S5).41 In the region between 1000 and 1300 cm-1, two

216

characteristic peaks typical of C-O functionality, are observed. The band at 1050-1100 cm-1 is

217

assigned to C-O-C groups (epoxy) groups, as it exists for GO and CGO-200, and subsequently

218

reduces for the rest samples. Adsorption at 1250 cm-1 is likely from –C-OH groups as it appears as

219

strong peaks for CGOs synthesized at ≥ 400 °C. Taken together, the data indicates that thermal

220

reduction initially starts with the removal of basal plane functional groups (e.g., epoxy) and then

221

proceeds to more chemically stable carbonyl and carboxyl functionalities at the material edge(s),

222

which is also supported by previous observations of GO materials.26

223

Aggregation Kinetics. Early-stage aggregation kinetics of GO/CGOs were assessed by measuring

224

the initial rate of change for hydrodynamic diameters as a function of time via time-resolved

225

dynamic light scattering (TR-DLS). For these materials, particle-particle interaction behaviors are a

226

function of both electrostatic repulsion (VEDL, due to electrostatic double layer) and van der Waals

227

attraction forces (VvdW).42, 43 Solution ionic strength (IS) influences the electrostatic repulsion forces

228

by affecting the inverse Debye length (Debye length κ ∝ IS0.5), and at low IS (low κ) the

229

interactions are described as long-range with high repulsion between interacting particles.42 With

230

additional electrolyte, electrostatic repulsion is further suppressed, and particle aggregation takes

231

place, as shown in an example aggregation profile (Figure S6). With sufficient electrolyte present

232

(over the critical coagulation concentration (CCC)), the total interaction becomes completely

233

attractive, leading to the transition from reaction-limited aggregation (RLA) to diffusion-limited

234

aggregation regimes (DLA) (Figure S6).

235

The attachment efficiency (ɑ) (also known as the inverse stability ratio) at different electrolyte

236

concentrations is calculated by normalizing the aggregation rate constant to the rate constant

237

obtained under diffusion-limited (attractive, fast) conditions, and is used to index particle aqueous

238

stability (details in Materials and Methods section). Particle-particle attachment efficiencies were 10 ACS Paragon Plus Environment

Environmental Science & Technology

239

plotted as a function of electrolyte concentrations in Figure 2. Distinct reaction-limited and

240

diffusion-limited regimes are observed for GO and CGOs within the concentration ranges of NaCl

241

(0 - 250 mM, Figure 2a), CaCl2 (0 - 50 mM, Figure 2b) and MgCl2 (0 - 50 mM, Figure 2c),

242

indicating that colloidal behavior follows classic Derjaguin-Landau-Verwey-Overbeek (DLVO)

243

theory.27

Page 12 of 27

244

CCC values were determined from the intersection of extrapolated lines through the diffusion-

245

and reaction-limited regimes (Table 1). CCC values determined here for GO (68.7 mM NaCl, 1.57

246

mM CaCl2, and 1.91 mM MgCl2) are between the values recently reported by Chowdhury et al. (44

247

mM NaCl, 0.9 mM CaCl2 and 1.3 mM MgCl2)23 and Wu et al. (188 mM NaCl, 2.6 mM CaCl2 and

248

3.9 mM MgCl2),28 likely due to varied surface chemistries as discussed above. This highlights the

249

importance of correlating the physical and chemical properties of GO to accurately predict colloidal

250

behavior. In the presence of MgCl2, CCC values are higher than those of CaCl2, which is also

251

consistent with previous reports,23, 28 due to the relatively weaker tendency of Mg2+ compared to

252

Ca2+ to form cation bridges (with carboxyl groups).44, 45 According to the Schulze-Hardy rule, the

253

ratio between CaCl2 and NaCl CCC could be approximated as Z-6 for colloids with high negative ζ-

254

potentials, where Z is the valence of Ca2+ ions (Z = 2).46 In our study, the ratios of CaCl2 and NaCl

255

CCC values for GO and CGOs synthesized below 600 °C are between Z-5.00 and Z-5.45, which is in

256

relatively good agreement with the rule. In contrast, such ratios were found to be Z-3.86 and Z-3.59 for

257

CGO-600 and CGO-800, deviating from the Schulze-Hardy prediction (Table 1). We hypothesize

258

such deviation is due to low ζ-potentials of CGO-600 and CGO-800, which violates the assumption

259

of the Schulze-Hardy rule, namely, the surface potential needs to be sufficiently high and remain

260

constant.46 Similar observations were also obtained in the presence of MgCl2 (Table 1).

261 262

Comparing the CCC values, it is also observed that, despite reduction, CGO-200, 400 and 500 have higher or similar CCC values compared to GO. For example, the NaCl CCC increased from 11 ACS Paragon Plus Environment

Page 13 of 27

Environmental Science & Technology

263

68.7 mM of GO to 81.7 mM of CGO-200, and 73.9 mM of CGO-400. This can be attributed to the

264

crumpling of GO structures, which can reduce the π-π interaction between discrete sheets, resulting

265

in aggregation-resistance.20, 47, 48 In particular, CGO-200, which retained much of the original

266

surface chemistry, compared to GO (see material characterization results), is more aggregation-

267

resistant in the presence of NaCl, CaCl2, and MgCl2 (CCC values were 19%, 59% and 62% higher

268

respectively (Figure 3)). The increases of CCC values in the presence of divalent cations (59% and

269

62% for Ca2+ and Mg2+ respectively) were greater compared to that of monovalent cations (19%),

270

which is a result of bridging/crosslinking behavior(s) of Ca2+ and Mg2+ ions.28, 49 With further

271

reduction (at higher synthesis temperature), CCC values decrease for all systems with a sharp

272

decrease occurring over the temperature window from 400 and 600 °C. There is no significant

273

difference between CCC values of CGO-600 and CGO-800 samples (Figure 3).

274

Correlating ζ-potentials and CCC. The CCC is defined as the minimum concentration of

275

electrolyte required to induce the coagulation (aggregation) of a stable colloidal suspension and can

276

be interpreted theoretically by the DLVO theory, which considers the electrostatic repulsion force

277

and the van der Waals attraction force between two interacting particles.

278

Conventionally, the van der Waals interaction is determined by employing the volume

279

integration approach (Hamaker’s technique), and the electrostatic repulsion interaction is obtained

280

by solving the Poisson-Boltzmann equation. However, for a complicated particle such as CGO, the

281

exact mathematical solutions are difficult to precisely ascertain. Instead, here the Derjaguin

282

approximation could be used, which scales the flat-plate interaction energy per unit area to the

283

corresponding interaction energy between two curved surfaces. The characterization length scales

284

of CGO particles (diameter: hundred nm; surface roughness: dozens of nm10) are significantly larger

285

than the interaction distance (e.g., Debye length: a few nm), thus making the Derjaguin

12 ACS Paragon Plus Environment

Environmental Science & Technology

Page 14 of 27

286

approximation applicable. By employing the Derjaguin approximation, the DLVO interaction

287

energies were solved by Hsu and Kuo50 and applied in our analysis. The electrical potential energy between two spherical particles VEDL can be estimated:50

288 289

!"

=

#$(%&)'( )* + ,

- , .

[ 01ℎ $ (

3 4

)]678 (−# :) ×